[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: memory problem



On Wednesday 2008 December 31 19:06:39 Ron Johnson wrote:
> On 12/31/08 17:41, Douglas A. Tutty wrote:
> > On Wed, Dec 31, 2008 at 03:21:02PM -0800, Ken Teague wrote:
> >> Douglas A. Tutty wrote:
> >>  > The only add-ons I have are from debian packages.  Even without
> >>  > addons I
> >>>
> >>> get memory leaks.
> >>
> >> Report the bug.  These things don't get fixed until they're addressed.
> >
> > Well, there are already many memory-related bugs.  See esp 452706.  The
> > complainer was running out with more than 1 GB of memory.  The suggested
> > action was to try FF3 in experimental.
> >
> > Like I'd want to run an experimental web browser given the propensity of
> > them to have security issues.
>
> How can s/w that's out of beta be "experimental"?

The software might not be "experimental" but the Debian package is.  Debian 
packages are (normally) built from source on the buildd boxes.  These boxes 
might have different libraries than upstream is using.  Debian might use 
different configure (or equivalent) options than upstream uses to build their 
binaries.  Debian may move the path of configuration, data, or even binaries 
to follow Debian policy and developers or other upstream testers are unlikely 
to have tested with those changes applied.

Also, a number of DDs feel that, during a freeze, uploads to unstable should 
be limited.  I disagree, but I'm not a DD (or even DM) and haven't had to put 
up with any of the issues that may arise from uploading to unstable during a 
freeze.

For those interested, I think the main problems is if a new version is in 
unstable, but inappropriate for (frozen) testing and a bug needs to be fixed 
in the testing version, an upload has to be made to testing-proposed-updates.  
Also, having different versions in testing and unstable can prevent packages 
otherwise ready for testing from moving there, making *other* maintainers and 
DD do more work -- and during a freeze testing diverges from unstable more 
and more.

I think there are ways to lessen the problem, but I haven't thought about it 
deeply and do not know how much extra load a solution might put on the Debian 
infrastructure.  While the most popular two architectures (i386, amd64) have 
fairly healthy infrastructure, other Debian supported architectures are 
already having some issues.  (E.g. HPPA and ARM recently had infrastructure 
related outages causing security updates to be delayed.)
-- 
Boyd Stephen Smith Jr.                     ,= ,-_-. =. 
bss@iguanasuicide.net                     ((_/)o o(\_))
ICQ: 514984 YM/AIM: DaTwinkDaddy           `-'(. .)`-' 
http://iguanasuicide.net/                      \_/     

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: